Analysis of Receptor for Vibrio cholerae El Tor Hemolysin with a Monoclonal Antibody That Recognizes Glycophorin B of Human Erythrocyte Membrane

Abstract

El Tor hemolysin (ETH), a pore-forming toxin secreted by Vibrio cholerae O1 biotype El Tor and most Vibrio cholerae non-O1 isolates, is able to lyse erythrocytes and other mammalian cells. To study the receptor for this toxin or the related molecule(s) on erythrocyte, we first isolated a monoclonal antibody, B1, against human erythrocyte membrane, which not only blocks the binding of ETH to human erythrocyte but also inhibits the hemolytic activity of ETH. Biochemical characterization and immunoblotting revealed that this antibody recognized an epitope on the extracellular domain of glycophorin B, a sialoglycoprotein of erythrocyte membrane. Erythrocytes lacking glycophorin B but not glycophorin A were less sensitive to the toxin than were normal human erythrocytes. These results indicate that glycophorin B is a receptor for ETH or at least an associated molecule of the receptor for ETH on human erythrocytes.

El Tor hemolysin/cytolysin (ETH) is a pore-forming toxin elaborated by most Vibrio cholerae non-O1 isolates (4, 18) and by Vibrio cholerae O1 El Tor isolates (7). This protein toxin is secreted as a 79-kDa protoxin that is proteolytically cleaved to yield an active toxin with a molecular mass of 65 kDa (26). It has been reported that the proregion of this toxin functions as a chaperone in the secretion process (17). Besides causing the hemolysis of various vertebrate erythrocytes, ETH induces the lysis of other mammalian cells and exhibits enterotoxicity in experimental diarrhea models (9). Thus, ETH may contribute to the pathogenesis of gastroenteritis caused by V. cholerae strains, especially the strains not producing cholera toxin, the major virulence factor (17).

ETH is believed to damage erythrocytes by acting as a pore-forming toxin with a series of possible processes. ETH first binds as a monomer to target cells, and then it assembles into detergent-stable oligomers and forms pores on the membrane of cells, and finally it causes colloidal osmotic lysis of cells (10, 27).

Although cholesterol is suggested to be essential for the oligomerization step (10), little is known about the mode of binding of ETH to the target cell. To study this binding step, especially the ETH receptor on human erythrocyte (one of the target cells most sensitive to ETH) (9), we immunized mice with human erythrocyte membrane and tried to develop monoclonal antibodies (MAb) against the ETH receptor or related molecule(s). By this process we obtained a MAb, B1, which can specifically inhibit the binding of ETH to the erythrocyte and further inhibit the hemolytic activity of ETH and used MAb B1 to determine the erythrocyte molecules associated with the receptor for ETH.

MATERIALS AND METHODS

Erythrocytes.

The normal human erythrocytes used in most of the experiments were from healthy volunteers. Erythrocytes having phenotypes of En(a−) cells, which lacked glycophorin A (5), and S-s-U− cells, which lacked glycophorin B (3), were also used. Cells deficient in these glycophorins were determined by biochemical and immunological analyses (3, 4) and were provided by Osaka Red Cross Blood Center.

Purification of ETH and preparation of antiserum against ETH.

ETH was purified from the culture supernatant of V. cholerae N86, as previously reported (24). Antiserum against ETH was prepared in rabbits inoculated intramuscularly with the purified ETH (100 μg) mixed with Freund adjuvant as previously described (25).

Production and purification of MAbs.

Immunization and the establishment of hybridoma against human erythrocytes membrane ghost prepared by hypotonic lysis with 5PB (5 mM sodium phosphate buffer, pH 8.0) (20) were essentially as previously described (8, 11, 12). Hybridomas selected in HAT (hypoxanthine-aminopterin-thymidine) medium (11, 12) were screened by enzyme-linked immunosorbent assay on plates coated with sonicated human erythrocyte membrane. The positive hybridomas were finally recloned at least twice by the limiting dilution method. The cloned hybridoma cells were injected into the peritoneum of BALB/c mice pretreated with 2,6,10,14-tetramethylpentadecane, and about 2 weeks later, ascitic fluid was collected. The antibody was purified from the collected ascitic fluid by using an Econo-Pac protein A kit (Bio-Rad). MAb against cholera toxin was prepared as described previously (13).

Screening of the MAbs that can inhibit the ETH hemolytic activity.

MAbs or their Fab fragments, prepared by using the ImmunoPure Fab preparation kit (Pierce Chemical Company) according to the manufacturer’s manual, were serially diluted with phosphate-buffered saline (PBS; 140 mM NaCl, 10 mM phosphate buffer [pH 7.2]) and mixed with a 4% (vol/vol) erythrocyte suspension at an equal volume (100 μl). The reaction mixtures were kept at 37°C for 30 min. After centrifugation at 2,000 × g for 2 min, the supernatant containing the extra unbound antibody was removed and then washed three times with 200 μl of PBS and reconstituted with PBS to 100 μl. Subsequently, 100 μl of ETH (final concentration, 0.1 μg/ml) was applied to every sample, and the mixtures were incubated at 37°C for 30 min. Then hemolytic activity was measured by a method previously described (25); briefly, after centrifugation at 2,000 × g for 2 min, 180 μl of supernatant of the reaction mixture was transferred to a 96-well plate for spectrophotometric measurement at 540 nm with the Multiscan MCC/340 (Labsystems, Tokyo, Japan).

Assay of MAb for inhibition of the binding of ETH to human erythrocytes by immunoblotting and immunocytochemistry.

Four percent human erythrocyte membrane ghost (100 μl) prepared as described above was incubated with MAb B1 (final concentration, 50 μg/ml) at 37°C for 30 min. For removing the extra antibody, the reacted erythrocyte membrane ghost was washed three times with PBS by aspirating the supernatant after centrifugation at 10,000 × g for 5 min. The erythrocyte membrane ghost was suspended to 100 μl by PBS and reacted with ETH (final concentration, 0.1 μg/ml) at 37°C for 5 min. After centrifugation at 10,000 × g for 5 min, the supernatant containing ETH unbound to erythrocyte membrane ghost was aspirated out, and the membrane ghost was washed five times with PBS. Then, the pellet of washed membrane was dissolved with 30 μl of Laemmli sample buffer but not heated, and 15 μl of the sample was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (14) with a separating gel containing 8% acrylamide. Control erythrocyte membrane ghost was incubated with PBS without MAb and reacted with ETH. The binding of ETH to the erythrocyte membrane ghost was determined by immunoblotting (22). Gel-electrophoresed sample was blotted onto a nitrocellulose transfer membrane (Protran; Schleicher & Schuell) and was probed with a 1:1,000 dilution of anti-ETH serum prepared as described above, followed with horseradish peroxidase (HRP)-F(ab′)₂-goat anti-rabbit immunoglobulin G (IgG) (H+L) (Zymed). Then the signal was developed with 4-chloro-1-naphthol (Wako Pure Chemical Industries, Ltd., Osaka, Japan).

Human erythrocytes were centrifuged at 8,000 rpm onto glass slides by using cytospin. The cells were then fixed with acetone-methanol (1:1 [vol/vol]) at 4°C for 10 min. Followed by three washes in PBS, the erythrocytes were pretreated with MAb (50 μg/ml) diluted with PBS at 37°C for 30 min. Unbound antibody was washed out by using PBS, and then the cells were incubated with ETH (0.1 μg/ml) at 37°C for 10 min. Finally, the extra ETH was washed out by PBS and an immunocytochemistry study was performed as described in a manual provided with the Histonfine SAB-AP(R) kit (Nichirei Corporation, Tokyo, Japan). As a positive control, the erythrocytes were treated with PBS instead of with MAb B1, and as a negative control, PBS was used instead of ETH.

Characterization of the erythrocyte membrane molecule recognized by MAb B1.

The following experiments were conducted to ascertain the nature of the erythrocyte membrane molecule that was recognized by MAb B1. (i) Human erythrocytes (4%) (200 μl) were treated for 1 h at 37°C with 1 mg of proteinase K from Trichiratum albus (Boehringer Mannheim)/ml or 50 μl of neuraminidase from Arthrobacter ureafaciens (Sigma)/ml. As a control, the same amount of erythrocytes untreated with the enzyme was used. After five washes by centrifugation for 3 min at 200 × g with PBS, the erythrocyte membrane ghost was prepared by lysing with 5PB as described above. This membrane ghost was separated by SDS–8% PAGE and blotted to a nitrocellulose membrane for immunoblotting, as described above, but probed with MAb B1 diluted 1:500. (ii) On the other hand, human erythrocytes (4%) (200 μl) were treated with 10 mM sodium metaperiodate–50 mM sodium acetate (pH 5.5) for 10 min at 0°C to selectively cleave carbons 8 and 9 of the unsubstituted side chain of terminal sialic acid (15) or with 10 mM sodium metaperiodate-50 mM sodium acetate (pH 4.5) for 1 h at room temperature to cleave carbon-carbon bonds between vicinal hydroxyl groups in (most) carbohydrates with a free carbon in position 3 (23). Controls were incubated with 50 mM sodium acetate buffer alone. After being washed with PBS, the cells were reduced by adding 50 mM sodium borohydride prepared in PBS (pH 7.6). Following three washes by centrifugation with PBS, the erythrocytes were lysed to prepare the membrane ghost, and the membrane ghost was used for immunoblotting analysis as described above. (iii) Two sets of normal human erythrocyte membranes ghost and two variant erythrocyte membranes ghost [En(a−) and S-s-U− erythrocytes] lacking glycophorins A and B, respectively, were resolved in SDS–8% PAGE and then blotted to a nitrocellulose transfer membrane for immunoblotting. One set was probed with a 1:500 dilution of MAb B1 as described above. The other set was probed with a 1:500 dilution of E4, a MAb against glycophorin A obtained from Sigma, followed with alkaline phosphatase-conjugated anti-mouse IgM (Sigma), and the signal was visualized by CDP-Star chemiluminescent substrate (Biolabs) diluted 1:250. (iv) Flow cytometric analysis was also used. Briefly, 4% various erythrocytes (100 μl) were allowed to interact with 100 μl of MAb B1 (final concentration, 50 pg/ml) at 37°C for 5 min and then washed with 500 μl of PBS three times by centrifugation. MAb-bound erythrocytes were then allowed to interact with a 1:1,000 dilution of fluorescein isothiocyanate (FITC)-labeled anti-mouse secondary antibodies (Cappel) and incubated for 1 h at 4°C. After being washed as described above, erythrocytes were suspended in 2 ml of PBS and loaded on FACScan (Becton Dickinson, San Jose, Calif.). The fluorescence intensity of FITC on 50,000 erythrocytes was recorded at 515 to 545 nm.

Comparison of the sensitivity of variant erythrocytes to ETH.

Four percent (vol/vol) cell suspensions of normal human erythrocytes, En(a−) erythrocytes, and S-s-U− erythrocytes were prepared with PBS, and 100 μl of these suspensions was mixed with 100 μl of a series-diluted ETH and incubated at 37°C for 30 min. After centrifugation at 2,000 × g for 2 min, the supernatant (180 μl) was used for the analysis of hemolytic activity as described above. One hundred percent hemolysis is defined as the optical density at 540 nm (OD₅₄₀) of hemoglobin released from erythrocytes that have been completely lysed by 0.1% Triton X-100 (Wako Pure Chemical Industries, Ltd., Osaka, Japan).

Other study methods.

Isotype determination of MAb was performed by using a Mono-Ab IFELA kit (Zymed Laboratories) according to the manufacturer’s instructions. Protein concentration was determined by using the Protein Assay Reagent kit (Pierce Chemical Company). MAb against cholera toxin was prepared as described previously (13).

RESULTS

Isolation of an MAb that inhibits the hemolytic activity of ETH on human erythrocytes.

In order to obtain MAb that recognizes the ETH receptor or related molecule(s), we immunized mice with human erythrocyte membrane ghost. As a result, a clone that secreted IgG class antibody, designated MAb B1, exhibited inhibition of hemolytic activity due to ETH in a dose-dependent manner (Fig. 1) but not thermostable direct hemolysin of Vibrio parahaemolyticus (data not shown). Fab fragment of MAb B1, however, could not inhibit the hemolytic activity of ETH as MAb B1 (data not shown). An unrelated MAb (MAb against cholera toxin) exhibited no inhibitory activity. Thus, MAb B1 specifically inhibits the hemolytic activity by ETH on human erythrocytes.

FIG. 1.

Inhibition of ETH-induced hemolytic activity by MAb B1. Erythrocytes were pretreated with various amounts of MAb B1 (●) and control (anticholera toxin) MAb (▴) before incubation with ETH. The ETH-induced hemolytic activity was indicated as the amount of released hemoglobin measured at OD₅₄₀. The axis values were expressed as percentages of hemolysis compared with the value in the condition without antibody (100% hemolysis). Each value is the mean of five experiments.

MAb B1 blocks the binding of ETH to human erythrocyte membrane.

The effect of MAb B1 on the inhibition of binding of ETH to human erythrocyte membrane was investigated by immunoblotting probed with antiserum to ETH. As shown in Fig. 2, no ETH signal was detected on human erythrocyte membrane pretreated with MAb B1, whereas the bound ETH—both the monomeric and the oligomeric forms—was clearly observed on erythrocytes without pretreatment of MAb B1, as described previously (10). The inhibition of ETH binding to human erythrocyte by MAb B1 was also confirmed by immunocytochemistry (data not shown). Thus, MAb B1 is able to block the binding of ETH to human erythrocyte and consequently to inhibit ETH-induced hemolysis.

FIG. 2.

Inhibition of ETH binding to human erythrocyte membrane. Erythrocyte membrane pretreated with MAb B1 (lane 2) or left untreated (lane 3) was incubated with ETH. After SDS-PAGE, the bound ETH was analyzed by immunoblotting by using anti-ETH serum. The purified ETH (10 pg) was subjected to immunoblotting with the sample as above (lane 1).

MAb B1 recognizes the sialylated glycoproteins on the human erythrocytes, and these sialoglycoproteins are glycophorin B.

The molecule of human erythrocyte membrane recognized by MAb B1 was analyzed by immunoblotting. As shown in Fig. 3, we observed that 46- and 25-kDa molecules of human erythrocyte membrane were recognized by MAb B1. However, after pretreatment of the erythrocytes with proteinase K to cleave the peptides or neuraminidase to liberate the sialic acid on the surface of cells, these two bands were not observed by immunoblotting with MAb B1. Furthermore, the human erythrocytes pretreated with sodium metaperiodate at pH 4.5 or 5.5 to destroy the sialic acid or carbon bonds (15, 23), respectively, also caused the loss of ability to bind with MAb B1. These results indicated that both the 46- and 25-kDa molecules recognized by MAb B1 are sialylated glycoprotein on the surface of human erythrocytes.

FIG. 3.

Characterization of the molecule recognized by MAb B1 as sialoglycoprotein. Equal amounts of erythrocytes were treated with proteinase K, neuraminidase, or periodate as described in Materials and Methods. Controls without enzyme or periodate were carried out in parallel (untreated lane). Then the membrane ghosts from these erythrocytes were subjected to immunoblotting with MAb B1.

Since it has been reported that glycophorin is the major sialoglycoprotein on the surface of erythrocyte membrane, and several molecular forms of glycophorin are known (2), the sialylated glycoprotein recognized by MAb B1 is likely glycophorin A or B. To confirm this, the membranes from variant erythrocytes which lacked glycophorin A or B were used for immunoblotting, together with normal erythrocyte membrane. As shown in Fig. 4, the membrane from S-s-U− erythrocyte which lacks glycophorin B (3) was not probed with MAb B1. However, the erythrocyte membrane from En(a−) erythrocyte, which lacks glycophorin A (5), was recognized by MAb B1 in the molecules of 46- and 25-kDa in molecular mass, identical to the pattern of normal human erythrocyte membrane. Furthermore, E4 (an MAb against glycophorin A; Sigma) recognized 37- and 70- to 83-kDa molecules on the normal erythrocytes or S-s-U− erythrocytes, but not on En(a−) erythrocytes. By flow cytometry (Fig. 5), we further observed that MAb B1 bound to En(a−) and normal human erythrocytes detected by FITC-labeled secondary antibodies (antimouse). However, the signal of MAb B1 was not detected on the surface of S-s-U− erythrocytes, which lack glycophorin B, like the negative control erythrocytes, which were not treated with MAb B1. Based on these findings and on the finding that glycophorin B was also observed as a dimer and monomer on the SDS-PAGE gel (1, 21), we believe that the sialoglycoprotein recognized by MAb B1 is glycophorin B and presents in dimeric and monomeric forms of 46 and 25 kDa, respectively, as shown by the immunoblotting analysis.

FIG. 4.

Analysis of erythrocyte membranes by immunoblotting with MAb B1 or E4. Membranes from S-s-U− erythrocyte (lane 1), En(a−) erythrocyte (lane 2), and normal human erythrocyte (lane 3) were separated by SDS-PAGE and blotted with MAb B1 or E4 (an MAb against glycophorin A).

FIG. 5.

Flow cytometric analysis of the molecule recognized by MAb B1. The solid lines represent control cells, which were not treated with MAb B1, and the dotted lines represent the sample cells, which were treated with MAb B1.

Erythrocytes that lack glycophorin B are less sensitive to ETH.

To ascertain whether there is a relation between glycophorin B and ETH-induced hemolysis, we determined the concentration of ETH required for 50% lysis in normal erythrocytes, En(a−) erythrocytes, and S-s-U− erythrocytes. As shown in Fig. 6, the concentration of ETH required for 50% hemolysis in normal human erythrocytes and En(a−) human erythrocytes, which lack glycophorin A, was about 60 pg/ml. However, about 650 pg of ETH/ml was required for 50% S-s-U− erythrocyte hemolysis. This indicates that the S-s-U− erythrocytes which lack glycophorin B are about 10 times less sensitive to ETH than En(a−) and normal human erythrocytes.

FIG. 6.

Comparison of the sensitivity of erythrocytes to ETH. A series of concentrations of ETH (100 μl) were incubated with the same volume of normal human erythrocyte (●), S-s-U− erythrocyte (■), and En(a−) erythrocyte (▴) at 37°C for 30 min. The hemolytic degree was determined as described in Materials and Methods. The data are expressed as a percentage of the amount of hemoglobin released from erythrocytes completely lysed by 0.1% Triton X-100. ETH concentrations for 50% hemolysis of various erythrocytes were compared.

DISCUSSION

In this study we first developed an MAb, B1, by immunization with the membrane of human erythrocyte, one of the cells that is most sensitive to ETH (24). This antibody inhibited ETH-induced hemolysis, as shown in Fig. 1. Thus, we investigated the mechanism of this inhibitory activity and found that MAb B1 inhibited the binding of ETH to the erythrocyte membrane receptor (Fig. 2), resulting in the inhibition of ETH-induced hemolysis. On the basis of these results, we proposed that the molecule(s) recognized by MAb B1 is associated with the functional binding (i.e., receptor) of ETH which leads to hemolysis.

Based on the result that treatment of erythrocytes with protease, neuraminidase, or periodate to cleave the peptides, sialic acid, or carbohydrates on the surface of erythrocytes leads to loss of ability to recognize MAb B1, we hypothesized that the molecule(s) recognized by MAb B1 is a sialoglycoprotein(s). Moreover, the results of immunoblotting with the membranes of erythrocytes which lacked glycophorin A or B [En(a−) or S-s-U−, respectively] confirmed that MAb B1 recognizes glycophorin B but not glycophorin A. Flow cytometric analysis gave the same result. Thus, it appears that MAb B1 recognizes the extracellular domain of glycophorin B.

Glycophorin B, the molecule recognized by MAb B1, is thus suggested to be the molecule associated with the receptor for ETH on human erythrocytes, but the receptor for ETH may or may not be identical to the epitope for MAb B1, and there is a possibility that the binding of MAb to glycophorin B sterically hinders the binding of ETH to the receptor. This possibility is supported by the fact that the Fab fragment of MAb B1 could not inhibit the hemolytic activity of ETH. Taken together, these results suggest that there is an interaction between ETH and glycophorin B, but the actual binding site for ETH on glycophorin B is not identical to the epitope for MAb B1.

We cannot presently identify the epitope(s) for ETH and MAb B1, because purification of glycophorin B is very complex (1) and purified glycophorin B is not available as a commercial product. Thus, we determined the sensitivity to ETH of human normal erythrocytes, S-s-U− erythrocytes, which lacked glycophorin B, and En(a−) erythrocytes, which lacked glycophorin A—the latter two being from very rare blood types (19)—and observed that the S-s-U− erythrocytes were about 10-fold less sensitive to ETH than the normal and En(a−) erythrocytes. Thus, we consider the most likely possibility to be that glycophorin B’s monomer or dimer or a complex molecule containing glycophorin B is the receptor for ETH. Since the S-s-U− erythrocytes were less sensitive to ETH but not completely resistant to it, glycophorin B may act as an associated molecule of the receptor for ETH, i.e., one of the components of the receptor complex for ETH. It is also possible that there are two forms of binding ETH at lower and higher concentrations (28). Glycophorin B is probably a receptor for the binding of ETH only at lower concentrations, such as those analyzed in this study. Some researchers have reported that ETH could lyse liposomes, which lacked glycophorin B, with higher ETH concentrations (10, 27, 28). We do not know the details about the mechanism of lysis induced by higher ETH concentrations. This mechanism may differ from that induced by lower ETH concentrations, and we are presently unable to speculate further.

Glycophorins A, B, C, and D constitute a group of erythrocyte transmembrane sialoglycoproteins (2). There is strong homology between the protein sequences of glycophorins A and B among these four sialoglycoproteins (1). Although glycophorin A has been an important molecule in the fields of membrane biochemistry and cellular biology for several decades (2, 6, 16, 19), the only biologic function ascribed to glycophorin B has been the carrying of blood group antigens Ss and U. This study is the first to suggest that glycophorin B is involved in ETH-induced hemolysis.

Furthermore, in our preliminary experiment MAb B1 also recognized the molecule(s) on the surface of intestinal ATCC 407 cells, which originated from the ileo-jejunum of a 2-month-old human embryo. Although the evidence in this study demonstrates that glycophorin B is a receptor molecule on human erythrocytes, this molecule (or immunologically cross-reactive molecule) may also be a receptor on intestinal cells. This interesting possibility should be addressed in future studies.